Off-line Text Independent Writer Identification Based On Texture Analysis

With the development of the technology of biometrics identification, the writer identification has gradually become a necessary part recently. It is an art of judging the identification of the writer by analyzing the same word characteristic or whole writing style of different people. It has been increasingly widely used in many areas, such as public security official, the acculturation examination and bank, etc. A typical application is the identification of the examinees in the higher education self-educated examination. Currently, these works are accomplished by hand, this kind of method has great workload, the low efficiency, and the accuracy of it is influenced by the subjective factor of person. Under this background, this paper synthesize the existing methods of writer identification, and then put forward a new method of writer identification using CDF 9/7 wavelet transform based on lifting scheme.Analyses the necessity and possibility of the writer identification, and introduces the concept of handwriting identification. Then integrates the research circumstance of domestic and international concerning the off-line writer identification, carries on a classification to the writer identification, and elaborates the present condition of writer identification, puts forward the project of carrying out writer identification system.Texture analysis is applied in the manipulation, analysis and recognition of images widely. The thesis summarizes many methods that were proposed in past several decades, in which wavelet transform has favorable time-frequency localization, scale transform and direction characteristic, and it has been a convincing tool for texture analysis as it can adapt human vision system much easily.This paper regards the feature of the handwriting picture as texture picture, creatively put forward the method of analyzing the handwriting picture with CDF 9/7 wavelet transform based on lifting, that is analyzing texture pictures with wavelet 9/7 on lifting scheme and denoting the feature vectors of the handwriting picture with the wavelet coefficient of the texture picture. Wavelet 9/7 on lifting scheme can extract the high frequency components of images effectively, high frequency components are the details of images, and analysis of texture images is mainly for the details. Experiments were made using Chinese handwriting from 23 different people, the contents are the same, and everyone has 5 copies. Obtain a good result In the environment of Windows2000,VC6.0, the identification ratio can reach 91.30% defending on the text.